Indonesian Journal of Electrical Engineering and Informatics (IJEEI)
Not a member yet
776 research outputs found
Sort by
Obstacle Evasion Algorithm Using Convolutional Neural Networks and Kinect-V1
Get PDFThe following paper presents the development of an algorithm for the evasion of static obstacles during the process of gripping the desired object, using an anthropomorphic robot, artificial intelligence, and machine vision systems. The algorithm has developed to detect a variable number of obstacles (between 1 and 15) and the grip desired element, using a robot with 3 degrees of freedom (DoF). A Kinect V1 was used to capture the RGB-D information of the environment and Convolutional Neural Networks for the detection and classification of each element. The capture of the three-dimensional information of the detected objects allows comparing the distance between the obstacles and the robot, to make decisions regarding the movement of the gripper to evade elements present in the path and hold the desired object without colliding. Obstacles of less than 18 cm in height were avoided, concerning the ground, with a probability of collision of 0% under specific environmental conditions, moving the robot since initial path in a straight line to the desired object, which is prone to changes according to the obstacles present in its. Function tests have been according to the manipulator's ability to evade possible obstacles of different heights located between the robot and the desired objec
Evaluation of Double Loop IC Module for Inductive Coupled Fed RFID Tag Wire Embedded Antenna
Get PDFThis paper presents an evaluation on double loop integrated circuit module for inductive couple fed RFID Tag. The inductive couple feed gives advantage especially for flexible wire embedded antenna in controlling the manufacturing varians as the chip is not directly connected to the antenna. Thus electrostatic discharged effect and manufacturing errors can be minimized. An inductor loop is directly connected to the RFID chip before it is placed to indirectly coupled the energy from a meandered dipole antenna. The coupled energy should be enough to turn on the chip to communicate with the reader. The inductive coupled energy is achieved by manipulating the near field magnetic field between the antenna body and the inductor loop. To evaluate the performance, the antenna and the inductive feeding loop is designed to operate at RFID UHF band (860 MHz – 960 MHz) and simulated using CST software. The antenna body impedance is evaluated to match the impedance of the chip and the loop. It is confirmed that the double loop inductor has higher inductance values and thus should be counted in conjugate impedance between the antenna body and the chip module. The tag with overall dimension of 60 mm x 16 mm can be read at distance at least 9 meters through out the band
The Prediction of Earthquake Building Structure Strength: Modified K-Nearest Neighbour Employment
Get PDFThe earthquake damage brings significant effects. The resilience of buildings against the earthquake and the destruction’s location is not an efficient outcome from previous research. This study applied the Modified K-Nearest Neighbor (MK-NN) in predicting the concrete structures’ performance despite the earthquakes. The 2-story building prediction covered earthquake history, time, concrete quality, displacement, velocity, and acceleration. The analysis of MK-NN provided the values of Euclidean, distance calculation, validity, and weight voting towards the classification of damages as “Safe” or “Immediate Occupancy” (IO). The K values exploited were 1, 3, 5, 7, 9, and 11, and simulation data training at 10:90, 20:80, 30:70. This study revealed the highest degree of accuracy at 98.85% with K=1 and a ratio of 30:70. Simultaneously, the lowest error rate was 1.15% at a similar K value and ratio. Herein, MK-NN significantly exceeds the accuracy and error rate of KNN up to 1.02% and 0.69%, respectively. To date, the automatic calculation prototyping software was then successfully developed. Ensuring the application’s accuracy, the Confusion Matrix, the Black box, and User Acceptance Test (UAT) have been performed. In a nutshell, this study provides a significant contribution to planning and information analysis of earthquake-resistant construction
Automated Learning of Hungarian Morphology for Inflection Generation and Morphological Analysis
Get PDFThe automated learning of morphological features of highly agglutinative languages is an important research area for both machine learning and computational linguistics. In this paper we present a novel morphology model that can solve the inflection generation and morphological analysis problems, managing all the affix types of the target language. The proposed model can be taught using (word, lemma, morphosyntactic tags) triples. From this training data, it can deduce word pairs for each affix type of the target language, and learn the transformation rules of these affix types using our previously published, lower-level morphology model called ASTRA. Since ASTRA can only handle a single affix type, a separate model instance is built for every affix type of the target language. Besides learning the transformation rules of all the necessary affix types, the proposed model also calculates the conditional probabilities of the affix type chains using relative frequencies, and stores the valid lemmas and their parts of speech. With these pieces of information, it can generate the inflected form of input lemmas based on a set of affix types, and analyze input inflected word forms. For evaluation, we use Hungarian data sets and compare the accuracy of the proposed model with that of state of the art morphology models published by SIGMORPHON, including the Helsinki (2016), UF and UTNII (2017), Hamburg, IITBHU and MSU (2018) models. The test results show that using a training data set consisting of up to 100 thousand random training items, our proposed model outperforms all the other examined models, reaching an accuracy of 98% in case of random input words that were not part of the training data. Using the high-resource data sets for the Hungarian language published by SIGMORPHON, the proposed model achieves an accuracy of about 95-98%
High-Performance Design of a 4-Bit Carry Look-Ahead Adder in Static CMOS Logic
Get PDFDesign of a 4-bit Carry Look-Ahead (CLA) process in static CMOS logic has been presented. CLA architecture proposed in this work computes carry-out terms without using carry-propagate and carry-generate signals which are used in conventional static CMOS (C-CMOS) 4-bit CLA adder. Performance parameters of the proposed 4-bit CLA architecture have been simulated and validated by comparing with the conventional design using Cadence design toolset in 45 nm technology. The designs were compared in terms of average power consumption, propagation delay and power delay product (PDP). The proposed 4-bit CLA topology obtained 34.53 % improvement in speed, 4.84 % improvement in power consumption and 37.696 % improvement in PDP while the source voltage was 1.0 V. Hence, as per acquired simulation results, the proposed 4-bit CLA structure is proven to be an excellent alternative to the conventional design for data-path design in modern high-performance processors.Design of a 4-bit Carry Look-Ahead (CLA) process in static CMOS logic has been presented. CLA architecture proposed in this work computes carry-out terms without using carry-propagate and carry-generate signals which are used in conventional static CMOS (C-CMOS) 4-bit CLA adder. Performance parameters of the proposed 4-bit CLA architecture has been simulated and validated by comparing with the conventional design using Cadence design toolset in 45 nm technology. The designs were compared in terms of average power consumption, propagation delay and power delay product (PDP). The proposed 4-bit CLA topology obtained 26.67 % improvement in speed, 5.966 % improvement in power consumption and 31.06 % improvement in PDP while the source voltage was 1.0 V. Hence, as per acquired simulation results, the proposed 4-bit CLA structure is proven to be an excellent alternative to the conventional design for data-path design in modern high-performance processors
Selecting Root Exploit Features Using Flying Animal-Inspired Decision
Get PDFMalware is an application that executes malicious activities to a computer system, including mobile devices. Root exploit brings more damages among all types of malware because it is able to run in stealthy mode. It compromises the nucleus of the operating system known as kernel to bypass the Android security mechanisms. Once it attacks and resides in the kernel, it is able to install other possible types of malware to the Android devices. In order to detect root exploit, it is important to investigate its features to assist machine learning to predict it accurately. This study proposes flying animal-inspired (1) bat, 2) firefly, and 3) bee) methods to search automatically the exclusive features, then utilizes these flying animal-inspired decision features to improve the machine learning prediction. Furthermore, a boosting method (Adaboost) boosts the multilayer perceptron (MLP) potential to a stronger classification. The evaluation jotted the best result is from bee search, which recorded 91.48 percent in accuracy, 82.2 percent in true positive rate, and 0.1 percent false positive rate
Active and Reactive Power Control of a Dual Stator Induction Machine (DSIM) using PI Controllers
Get PDFIn this paper, we present the active and reactive power control of a Dual Stator Induction Generator (DSIG). Contrary to the Doubly Fed Induction Generator (DFIG), which is controlled by its rotor and which delivers power from its stator, in the DSIG, one of the two stator windings plays the role of control winding and the other plays that of the power winding. The aim of this article is to establish, using some simplifying assumptions, the relationships between the active and the reactive powers delivered from the power winding and the voltages applied at its control winding terminals. Based on these relationships, a functional diagram is built and the active and reactive power regulators are synthesized. A model of the controlled DSIG is implemented under MATLAB-Simulink environment and the simulation results showed the feasibility and the performances of the developed control laws
Application of modified least squares method for order reduction of commensurate higher order fractional systems
Get PDFThe paper related to the reduction and investigation of family of commensurate fractional order systems. The fractional order system is first transformed to integer order and then a hybrid method is applied as a model reduction scheme. In this scheme the reduced denominator is acquired by least square method and the numerator is achieved by time moment matching. This formulated reduced integer model is reconverted to reduced fractional model. Three examples are conferred to authenticate and emphasize the efficacy of the proposed approach. The proposed method is verified by MATLAB simulation, and shows better performance in the estimation process
Double-talk robust acoustic echo canceller based on CNN filter
Get PDFConventional acoustic echo cancellation works by using an adaptive algorithm to identify the impulse response of the echo path. In this paper, we use the CNN neural network filter to remove the echo signal from the microphone input signal, so that only the speech signal is transmitted to the far-end. Using the neural network filter, weights are well converged by the general speech signal. Especially it shows the ability to perform stable operation without divergence even in the double-talk state, in which both parties speak simultaneously. As a result of simulation, this system showed superior performance and stable operation compared to the echo canceller of the adaptive filter structure
Recognition of Badminton Action Using Convolutional Neural Network
Get PDFDeep learning approach has becoming a research interest in action recognition application due to its ability to surpass the performance of conventional machine learning approaches. Convolutional Neural Network (CNN) is among the widely used architecture in most action recognition works. There are various models exist in CNN but no research has been done to analyse which model has the best performance in recognizing actions for badminton sport. Hence, in this paper we are comparing the performance of four different pre-trained models of deep CNN in classifying the badminton match images to recognize the different actions done by the athlete. Four models used for comparison are AlexNet, GoogleNet, VggNet-16 and VggNet-19. The images used in this experimental work are categorized into two classes: hit and non-hit action. Firstly, each image frame was extracted from Yonex All England Man Single Match 2017 broadcast video. Then, the image frames were fed as the input to each classifier model for classification. Finally, the performance of each classifier model was evaluated by plotting its performance accuracy in form of confusion matrix. The result shows that the GoogleNet model has the highest classification accuracy which is 87.5% compared to other models. In a conclusion, the pre-trained GoogleNet model is capable to be used in recognizing actions in badminton match which might be useful in badminton sport performance technology